A novel type of wireless device sends and receives data without a battery or other conventional power source. Instead, the devices harvest the energy they need from the radio waves that are all around us from TV, radio, and Wi-Fi broadcasts.
These seemingly impossible devices could lead to a slew of new uses of computing, from better contactless payments to the spread of small, cheap sensors just about everywhere.
“Traditionally wireless communication has been about devices that generate radio frequency signals,” says Shyam Gollakota, one of the University of Washington researchers who led the project. “But you have so many radio signals around you from TV, Wi-Fi, and cellular networks. Why not use them?”
Gollakota and colleagues have created several prototypes to test the idea of using ambient radio waves to communicate. In one test, two credit-card-sized devices—albeit with relatively bulky antennas attached—were used to show how the technique could enable new forms of payment technology. Pressing a button on one card caused it to connect with and transfer virtual money to a similar card, all without any battery or external power source.
Here is a video of the prototypes:
“In that demonstration, the LEDs, touch sensors, microcontrollers, and the wireless communication are all powered by those ambient TV signals,” says Gollakota.
The devices communicate by varying how much they reflect—a quality known as backscatter—and absorb TV signals. Each device has a simple dipole antenna with two identical halves, similar to a classic “rabbit ears” TV aerial antenna. The two halves are linked by a transistor, which can switch between two states. It either connects the halves so they can work together and efficiently absorb ambient signals, or it leaves the halves separate so they scatter rather than absorb the signals. Devices close to one another can detect whether the other is absorbing or scattering ambient TV signals. “If a device nearby is absorbing more efficiently, another will feel [the signals] a bit less; if not, then it will feel more,” says Gollakota. A device encodes data by switching between absorbing and not absorbing to create a binary pattern.
The device gets the power to run its electronics and embedded software from the trickle of energy scavenged whenever its antenna is set to absorb radio waves.
In the tests, the devices were able to transfer data at a rate of one kilobit per second, sufficient to share sensor readings, information required to verify a device’s identity, or other simple tidbits. So far the longest links made between devices are around 2.5 feet, but the University of Washington team could extend that to as much as 20 feet with some relatively straightforward upgrades to the prototypes. The researchers also say the antennas of backscatter devices could be made smaller than those in the prototypes.
Gollakota says the devices could be programmed to work together in networks in which data travels by hopping from device to device to cover long distances and eventually connect to nodes on the Internet. He imagines many of a person’s possessions and household items being part of that battery-free network, making it possible to easily find a lost item like your keys. “These devices can talk to each other and know where it is,” he says.
The researchers tested that scenario by placing tags on cereal boxes lined up on a shelf to mimic a grocery store or warehouse. Each tag communicated with its nearest neighbor to check if it was in the correct place, and blinked its LED if it was not.
That demonstration impresses Kristofer Pister, a professor at the University of California, Berkeley, whose work on tiny devices dubbed “smart dust,” which gather data from just about anywhere, helped spawn many research projects on networked sensors. Using TV signals to enable such applications without batteries is “a really clever idea,” he says.
While Pister and others around the world—including the Washington group—have spent years creating the technology needed to make cheap, compact sensors practical (see “Smart Specks”), such networks are relatively scarce. Josh Smith, a University of Washington professor who led the backscatter project with Gollakota, says that being able to do without onboard power could help.
“The need for batteries is one thing that has been slowing down their deployment,” he says. Without batteries, sensors can be significantly cheaper, and much longer-lasting, allowing them to be placed in areas otherwise not worth it, says Smith. “You could build sensors into the walls of a building knowing they would work years later.”
Bhaskar Krishnamachari, who works on sensor networks at the University of Southern California, notes that in some rural areas and indoor environments, there may not be enough ambient radio waves to support the battery-free approach. “For many practical implementations, an onboard battery may be unavoidable,” he says. “However, the proposed approach may go some way in extending the time between battery-charging events.”
The backscatter communication technology was developed by Gollakota with Smith and David Wetherall, also a University of Washington professor, along with grad students Vincent Liu, Aaron Parks, and Vamsi Talla. A paper on the technology won best paper award at the ACM Sigcomm conference in Hong Kong this week.